JPH04170599A - Text voice synthesizing device - Google Patents

Abstract

PURPOSE:To transmit information accurately further plainly by providing a homophone discriminating mode select means having a normal mode of outputting a voice through normal read and a homophone disriminating mode of inserting read, for properly discriminating a part of a homophone in the case of providing the homophone, to output a voice. CONSTITUTION:There are provided a word processing part 1, voice synthesizing part 2 and a homophone discriminating mode select means having a normal mode of outputting a voice through normal read and a homophone discriminating mode of inserting read, for properly discriminating a part of a homophone in the case of providing the homophone, to output a voice. Thus by providing the homophone discriminating mode select means having the normal mode of outputting the voice through the normal read and the homophone discriminating mode of inserting the read, for properly discriminating the part of the homophone in the case of providing the homophone, to output the voice, information can be accurately further plainly transmitted even in the case of an information transmitting means by only a voice.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a text-to-speech synthesis device capable of editing phonetic symbol strings.

Conventional technology Conventionally, in text-to-speech synthesis devices, a language processing unit performs linguistic analysis on a coded character string and outputs a "phonetic symbol string" that includes reading, accent, intonation, and other voice output control information. , a speech synthesis unit interprets the phonetic symbol string and converts it into synthesized speech. Therefore, if you give it a coded character string as input, it will automatically output it as voice.

Problems to be Solved by the Invention However, in the case of the conventional text-to-speech synthesizer described above, apart from the case where the voice can be heard while visually viewing the text output through a display, only the voice is used as information. When using words as a means of communication, it is not possible to correctly recognize homophones (words that have the same pronunciation but different spellings and meanings) just by reading them normally.

Conventionally, there are devices that have a mode for identifying and reading such homophones for the purpose of proofreading texts or for blind people, but such modes are not designed for the purpose of identifying homophones or symbols. Therefore, when a general user uses the text-to-speech synthesizer as a voice information transmission tool, there is a problem in that it becomes difficult to grasp the content of the text.

Means for Solving the Problem Therefore, in order to solve such problems, claim 1
The invention described linguistically analyzes a coded character string,
a language processing unit that outputs a phonetic symbol string containing voice output control information such as normal reading, reading for identifying homophones, accent, intonation, etc.; and a speech synthesis unit that converts the phonetic symbol string into synthesized speech. , a homophone identification mode selection means having a normal mode in which audio is output in the normal reading, and a homophone identification mode in which when there is a homophone, the audio is output by inserting a pronunciation for correctly identifying the part of the homophone. Established.

The invention according to claim 2 is the invention according to claim 1,
The speech synthesis unit allows the user to freely specify the position of the output sentence, and also performs homophone identification speech output control that inserts pronunciations for identifying homophones only at predetermined positions specified by the user and outputs the speech. I prepared the means.

The invention according to claim 1 provides a normal mode in which a voice is output in the normal reading, and a homophone identification mode in which when there is a homophone, a reading is inserted to correctly identify the part of the homophone and the voice is output. By providing the homophone identification mode selection means, it is possible to transmit information accurately and in an easy-to-understand manner even when the information transmission means is only audio.

The invention as claimed in claim 2 provides a homophone identification voice in which a user can freely specify the position of an output sentence, and a pronunciation for identifying a homophone is inserted only at a predetermined position specified by the user and output as voice. By providing the output control means, it is possible to avoid the trouble of inserting pronunciations for identifying homophones one by one for homophones that can be identified from the context.

Embodiment An embodiment of the present invention will be explained based on FIGS. 1 to 4. As shown in FIG. 1, this device includes a language processing section 1, a speech synthesis section 2, and homophone identification mode selection means (not shown). In this case, the language processing unit 1
Linguistically analyzes the coded character string of the input sentence a and outputs a phonetic symbol string containing audio output control information such as normal reading, reading for identifying homophones, accent, and intonation. There is work. This language processing unit 1
A word dictionary 3 is connected to. Said speech synthesis section 2
has the function of converting the phonetic symbol sequence into synthesized speech C.

The homophone identification mode selection means has a "normal mode J" which outputs audio in normal pronunciation, and a "homonym J" which outputs audio by inserting a pronunciation for correctly identifying the homophone part when there is a homophone. identification mode.

Further, the speech synthesis section 2 is provided with homophone identification speech output control means (not shown). This homophone identification voice output control means allows the user to freely specify the position of the output sentence, and also has the function of inserting a pronunciation for identifying the homophone only at a predetermined position specified by the user and outputting the voice. There is.

In this configuration, the functions of the homophone identification mode selection means and the homophone identification voice output control means will be mainly described below.

It is assumed that a sentence a input to the language processing unit 1 is in Japanese. The language processing unit 1 refers to the word dictionary 3, performs linguistic analysis on the input sentence a, and outputs a string of pronunciation symbols including pronunciation, accent, intonation, etc. At this time, for words that have homophones, a pronunciation for identifying the homophone is registered in the word dictionary 3, and the pronunciation is inserted into the normal pronunciation during language analysis.

Here, the function of the homophone identification mode selection means will be explained based on an example of the word dictionary 3 in FIG. 2 and an example of the language processing result in FIG. 3. Now, let us assume that there is an input "sentence a" as shown in Fig. 3(a), and as shown in Fig. 3(a),
Assume that a "phonetic symbol string b" as shown in b) is obtained. At this time, full-width hiragana characters in the phonetic symbol string represent readings, and half-width symbols represent control codes. ``〈''', ``〉'' in the phonetic symbol string represent the start and end positions of pronunciations to identify homophones, respectively.In the ``normal mode'', the speech synthesis unit 2 (II, +1
> II (TI, II > T
The part surrounded by I is also output as audio. FIG. 4(a) shows an example of the output audio in the normal mode, and FIG. 4(b) shows an example of the output audio in the homophone identification mode. Also,
If you want to insert a pronunciation that identifies homophones only in the part specified by the user, insert “<”,
The part surrounded by ``>'' should also be output as audio.

As mentioned above, homophone identification has two modes: a normal mode in which audio is output using the normal reading, and a homophone identification mode in which when there is a homophone, a reading is inserted to correctly identify the part of the homophone and the audio is output. By providing the mode selection means, even if the information transmission means is only audio, it becomes possible to transmit information accurately and in an easy-to-understand manner.

Next, the function of the homophone identification speech output control means of the speech synthesis section 2 will be explained. As described above, the language processing unit 1
The phonetic symbol strings output as a result of language processing are stored in a recording medium (not shown). Then, the speech synthesis section 2
Then, the phonetic symbol sequence is read out from the storage medium as appropriate and converted into synthesized speech C. At this time, it is assumed that the speech synthesis unit 2 reads from the position of a "pointer" indicating the reading start position of the phonetic symbol string, and this pointer is incremented by the amount read. Furthermore, although the initial value of this pointer is set to indicate the beginning of sentence a, this position can be changed freely by the user. Therefore, by moving this pointer forward, it is possible to repeatedly output sentence a that has already been audio output, and by moving the pointer backward, it is possible to skip sentence a and proceed forward. can.

As mentioned above, the user can freely specify the position of the output text, and the homophone identification voice output control inserts a pronunciation for identifying homophones only at the predetermined position specified by the user and outputs the voice. By providing means,
For homophones that can be identified from the context, it is possible to avoid the trouble of inserting pronunciations to identify each homophone.

Effects of the Invention The invention as claimed in claim 1 analyzes coded character strings to produce phonetic symbols that include audio output control information such as normal pronunciation, pronunciation for identifying homophones, accent, and intonation. a language processing unit that outputs a string, a speech synthesis unit that converts the phonetic symbol string into synthesized speech, a normal mode that outputs speech in normal reading, and a normal mode that outputs speech in normal reading, and a system that correctly identifies the homophone part when there is a homophone. Since the homophone identification mode selection means has a homophone identification mode that inserts a pronunciation and outputs audio, it is possible to accurately and easily transmit information even when the information transmission means is audio only. It is possible.

The invention according to claim 2 is the invention according to claim 1,
The speech synthesis unit allows the user to freely specify the position of the output sentence, and also performs homophone identification speech output control that inserts pronunciations for identifying homophones only at predetermined positions specified by the user and outputs the speech. I have prepared the means,
For homophones that can be identified from the context, it is possible to avoid the trouble of inserting pronunciations to identify each homophone.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is an explanatory diagram showing an example of the word dictionary, and FIG. 3 is an explanatory diagram showing the result of language processing between an input sentence and a string of phonetic symbols. , 4th
The figure is an explanatory diagram showing an example of output audio in the normal mode and the isomorphic identification mode. 1... Phonetic symbol string, 2... Speech synthesis section, b... Phonetic symbol string Applicant Rico Co., Ltd. - He had a taste for science. (a) The input sentence (is) is <<(δ)isense no (ka)gaku>
niba kiyo) - mio (motsu). (b) In the evocative symbol sequence, he had great respect for the light (a) In normal mode, he had great respect for the light of the light (b) Isomorphism 1! identification mote

Claims (1)

[Claims] 1. A phonetic symbol string containing speech output control information such as normal pronunciation, pronunciation for identifying homophones, accent, intonation, etc. by linguistically analyzing a coded character string. a language processing unit that outputs the phonetic symbol string, a speech synthesis unit that converts the phonetic symbol string into synthesized speech, a normal mode that outputs speech in normal reading, and a reading that correctly identifies the homophone when there is a homophone. 1. A text-to-speech synthesis device comprising: a homophone identification mode selection means having a homophone identification mode for inserting a ``phonetic word'' and outputting a voice. 2. The speech synthesis unit allows the user to freely specify the position of the output sentence, and outputs homophone identification speech by inserting pronunciations for identifying homophones only at predetermined positions specified by the user. The text-to-speech synthesis apparatus according to claim 1, further comprising a control means.